A large slope inclined column splitting and accurate and rapid alignment method

Abstract

This invention relates to the field of steel structure construction technology and discloses a method for the disassembly and precise rapid alignment of inclined columns with large inclinations. The method includes: deploying a UWB base station cluster and a visual calibration board on-site to complete a global coordinate system; scanning completed foundation segments, reconstructing local BIM models, and exporting target pose parameters; transporting prefabricated inclined columns to their designated locations, affixing reflective markers, and loading an IMU module; slowly lifting the column after hooking it onto a crane while simultaneously activating the attitude estimator; switching to dynamic trajectory planning and active compensation control mode to perform fine approximation; activating the quick-release interface self-locking upon reaching a preset threshold, and releasing the hook after confirming RFID verification. This invention, through the deep integration of multi-source sensing technology, advanced control theory, and BIM digitalization, successfully transforms the hoisting of inclined columns with large inclinations from an "experience-dependent" to a "data-driven" method, significantly improving construction safety and efficiency while ensuring extremely high precision.

Description

Technical Field

[0001] This invention relates to the field of steel structure building construction technology, specifically a method for splitting and accurately and quickly aligning inclined columns with large slopes. Background Technology

[0002] In complex steel structure projects such as large stadiums, airport terminals, and super high-rise buildings, steeply inclined columns are key components that support the overall structural load. Due to the diverse requirements of architectural design, these inclined columns often have characteristics such as large cross-sectional changes, steep inclination angles, and long single-column lengths. Due to limitations imposed by road transport clearances, construction site conditions, and the lifting capacity of hoisting equipment, steeply inclined columns usually cannot be manufactured as a whole or hoisted in one go. They must be prefabricated (disassembled) in sections in the factory or on-site, and then transported to the site for high-altitude splicing and assembly.

[0003] However, existing technologies for segmented hoisting and alignment of inclined columns with high slopes have the following significant drawbacks and technical bottlenecks:

[0004] Traditional hoisting operations mainly rely on manual measurement with total station and the experience of crane operators. Under steep inclines, the center of gravity of the components shifts significantly, and long slings are susceptible to wind loads, resulting in a "pendulum effect" and high-frequency vibrations. Furthermore, manual operation is slow to react and cannot compensate for dynamic disturbances in real time, leading to alignment errors of single segments typically at the centimeter level. For ultra-long inclined columns with multiple segments, these errors accumulate layer by layer, easily causing the final total length deviation to exceed the standard or the straightness to fail to meet design requirements. In severe cases, it can even lead to failure to close the closure or the need for forced hole enlargement and flame straightening, seriously affecting structural safety. Summary of the Invention

[0005] This invention provides a method for splitting and accurately and quickly aligning inclined columns with high slopes. By deeply integrating multi-source sensing technology, advanced control theory and BIM digitalization, it successfully transforms the hoisting of inclined columns with high slopes from an "experience-dependent" to a "data-driven" method. While ensuring extremely high precision, it significantly improves construction safety and efficiency, and solves the problem mentioned in the background art where the accumulation of errors in hoisting with high slopes leads to difficulties in closure, seriously threatening construction safety and efficiency.

[0006] This invention provides the following technical solution:

[0007] A method for high-slope oblique column decomposition and precise and rapid alignment includes the following steps:

[0008] Step S1: Deploy the UWB base station cluster and visual calibration board on site to complete the global coordinate system one;

[0009] Step S2: Scan the completed basic segments, reconstruct the local BIM model, and export the target pose parameters;

[0010] Step S3: Transport the prefabricated inclined columns to their designated locations, affix reflective markings, and load the IMU module;

[0011] Step S4: After the crane hooks up, it is slowly lifted while the attitude estimator is activated simultaneously;

[0012] Step S5: Switch to dynamic trajectory planning and active compensation control mode to perform fine approximation;

[0013] Step S6: After reaching the preset threshold, activate the quick-release interface self-locking, and release the hook after confirming that the RFID verification is passed;

[0014] Step S7: Repeat the above process until the entire inclined column is assembled. Finally, check the total length error with a total station to ensure it is ≤ ±2mm.

[0015] As a preferred embodiment of the present invention, the specific process of completing the global coordinate system in step S1 includes:

[0016] S101. Using the visual calibration board as a common reference, collect the absolute coordinates of the UWB base station group in the global coordinate system and the relative coordinates in the visual camera coordinate system respectively.

[0017] S102. Calculate the rotation and translation transformation matrix between the UWB coordinate system and the visual coordinate system using the hand-eye calibration algorithm;

[0018] S103. Define the center point of the vision calibration board as the origin of the global coordinate system, map all sensor data to the global coordinate system, and eliminate the spatial reference deviation of multi-source heterogeneous data.

[0019] As a preferred technical solution of the present invention, the specific process of reconstructing the local BIM model and deriving the target pose parameters in step S2 includes:

[0020] S201. Denoise, register, and reconstruct the surface of the collected point cloud data to generate a high-precision three-dimensional mesh model of the basic segments.

[0021] S202. Iteratively register the generated 3D mesh model with the theoretical BIM model in the design phase, and calculate the deviation matrix between the actual installation state and the design state.

[0022] S203. Based on the deviation matrix correction theory, generate a generator containing position vectors from the coordinate values ​​of the docking surface. and attitude quaternions The target pose parameter set is obtained and sent to the control terminal.

[0023] As a preferred embodiment of the present invention, the operation logic of the attitude estimator in step S4 specifically includes:

[0024] S401. The angular velocity and acceleration output by the IMU module are used as prediction inputs, and the absolute position output by the UWB base station group and the relative pose calculated by the vision system from the reflective markers are used as observation inputs.

[0025] S402. Construct extended Kalman filter state equations to dynamically compensate for high-frequency vibration, wind load disturbance, and elastic deformation of slings during the hoisting process of the inclined column.

[0026] S403 outputs six-degree-of-freedom attitude data in real time as feedback signals for crane motion control.

[0027] As a preferred embodiment of the present invention, in step S5, switching to the dynamic trajectory planning and active compensation control mode to perform fine approximation specifically includes:

[0028] S501. Based on the error vector between the current real-time attitude and the target pose parameters, construct a nonlinear model to predict the control objective function;

[0029] S502. Under the conditions of satisfying the maximum speed, maximum acceleration and anti-collision constraints of the crane, the control command sequence in the future time domain is optimized by rolling.

[0030] S503: The optimized control commands are decomposed into the crane's luffing, slewing, and hoisting actions, as well as the attitude fine-tuning commands of the inclined column itself, to achieve millimeter-level trajectory tracking;

[0031] Among them, fine approximation is 1m from the docking surface.

[0032] As a preferred embodiment of the present invention, the specific process of activating the quick-release interface self-locking in step S6 includes:

[0033] S601. When the vision system detects that the distance between the magnetic guide cone pin at the end of the inclined column segment and the reserved hole in the foundation segment is less than 50mm, the vision servo closed-loop guidance is activated.

[0034] S602. When the contact sensor detects a physical contact signal, it immediately triggers the pneumatic clamping ring to inflate and lock, generating an axial locking force greater than 50kN.

[0035] As a preferred embodiment of the present invention, in step S6, RFID verification specifically includes:

[0036] S603. Read the identification information of the RFID chip embedded at the end of the inclined column segment.

[0037] S604. Compare the installation sequence information with the preset information in the local BIM model.

[0038] S605: If they match, send an unlock command; if they do not match, issue an audible and visual alarm and lock the hook.

[0039] As a preferred embodiment of the present invention, step 7 of the total station verification includes the following specific implementation steps:

[0040] S701. After the entire inclined column is assembled, select at least three characteristic control points at the top, bottom and middle of the inclined column for total station coordinate measurement.

[0041] S702. Import the measured coordinate data into the data processing software and calculate the cumulative length deviation and overall straightness deviation between each segment.

[0042] S703. If the total length error exceeds ±2mm or the straightness deviation exceeds the design allowable value, the local fine-tuning program is initiated. The hydraulic jacking device is used to release stress and correct the position of specific segments until the accuracy requirements are met.

[0043] Compared with existing technologies, this invention provides a method for high-slope oblique column splitting and precise and rapid alignment, which has the following beneficial effects:

[0044] 1. This method for splitting and precisely aligning large-slope inclined columns, through multi-scale sensor fusion of "UWB wide-area positioning + IMU real-time inertial measurement + visual end-point guidance" and combined with model predictive control, controls the traditional centimeter-level hoisting error to a total length deviation of ≤±2mm, meeting the stringent requirements of precision splicing of ultra-large inclined columns, achieving "millimeter-level" extreme accuracy, and improving the alignment accuracy of large-slope inclined columns.

[0045] 2. This method for splitting and precisely aligning inclined columns with a large slope utilizes RFID identity verification and BIM sequence comparison to form a hard logic lock, fundamentally eliminating the risk of rework caused by incorrect component installation sequence; and provides an axial locking force of more than 50kN through a pneumatic clamping ring, which, together with the contact sensor trigger, ensures the stability of the connection at the moment of connection and prevents slippage accidents.

[0046] This method, by deeply integrating multi-source sensing technology, advanced control theory and BIM digitalization, successfully transforms the hoisting of inclined columns with large slopes from "experience-dependent" to "data-driven," significantly improving construction safety and efficiency while ensuring extremely high precision. Attached Figure Description

[0047] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, the elements or parts are not necessarily drawn to actual scale.

[0048] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0049] Figure 2 This is a schematic diagram of the system framework of the present invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Example:

[0052] Reference Figures 1-2 A method for high-slope oblique column decomposition and precise and rapid alignment includes the following steps:

[0053] Step S1: Deploy the UWB base station cluster and visual calibration board on site to complete the global coordinate system one;

[0054] Step S2: Scan the completed basic segments, reconstruct the local BIM model, and export the target pose parameters;

[0055] Step S3: Transport the prefabricated inclined columns to their designated locations, affix reflective markings, and load the IMU module;

[0056] Step S4: After the crane hooks up, it is slowly lifted while the attitude estimator is activated simultaneously;

[0057] Step S5: Switch to dynamic trajectory planning and active compensation control mode to perform fine approximation;

[0058] Step S6: After reaching the preset threshold, activate the quick-release interface self-locking, and release the hook after confirming that the RFID verification is passed;

[0059] Step S7: Repeat the above process until the entire inclined column is assembled. Finally, check the total length error with a total station to ensure it is ≤ ±2mm.

[0060] In step S1, the specific process of completing the global coordinate system includes:

[0061] S101. Using the visual calibration board as a common reference, collect the absolute coordinates of the UWB base station group in the global coordinate system and the relative coordinates in the visual camera coordinate system respectively.

[0062] S102. Calculate the rotation and translation transformation matrix between the UWB coordinate system and the visual coordinate system using the hand-eye calibration algorithm;

[0063] S103. Define the center point of the vision calibration board as the origin of the global coordinate system, map all sensor data to the global coordinate system, and eliminate the spatial reference deviation of multi-source heterogeneous data.

[0064] In step S2, the specific process of reconstructing the local BIM model and deriving the target pose parameters includes:

[0065] S201. Denoise, register, and reconstruct the surface of the collected point cloud data to generate a high-precision 3D mesh model of the basic segments.

[0066] Among them, the point cloud data is the three-dimensional coordinate information of different points on the inclined column;

[0067] S202. Iteratively register the generated 3D mesh model with the theoretical BIM model in the design phase, and calculate the deviation matrix between the actual installation state and the design state.

[0068] S203. Based on the deviation matrix correction theory, generate a generator containing position vectors from the coordinate values ​​of the docking surface. and attitude quaternions The target pose parameter set is obtained and sent to the control terminal.

[0069] in, Represents the three components in a three-dimensional Cartesian coordinate system. The planar coordinates representing the horizontal plane determine the specific point where the inclined column lands in the horizontal direction. The height represents the vertical height of the inclined column segment.

[0070] in, Represents the spatial rotation attitude. The component representing the axis of rotation (vector part), The component representing the rotation angle (scalar part).

[0071] In step S4, the operation logic of the attitude estimator specifically includes:

[0072] S401. The angular velocity and acceleration output by the IMU module are used as prediction inputs, and the absolute position output by the UWB base station group and the relative pose calculated by the vision system from the reflective markers are used as observation inputs.

[0073] S402. Construct extended Kalman filter state equations to dynamically compensate for high-frequency vibration, wind load disturbance, and elastic deformation of slings during the hoisting process of the inclined column.

[0074] S403 outputs six-degree-of-freedom attitude data in real time as feedback signals for crane motion control.

[0075] The specific description of the extended Kalman filter state equation is as follows:

[0076] In S402, the traditional rigid body kinematics model is insufficient to describe the dynamic characteristics of an inclined column under a long sling. Therefore, an augmented state vector is constructed, adding sling swing mode state variables to the original position, velocity, attitude, and zero bias: ;

[0077] The state transition equation introduces a damped simple harmonic oscillation term:

[0078]

[0079] in The natural frequency of the sling (g is the acceleration due to gravity), (Effective length of sling) For the air damping ratio, θ = θ swing The swing angle of the sling swing mode. = The first derivative of the swing angle is the angular velocity. The second derivative of the swing angle is the angular acceleration. This is an external disturbance input.

[0080] Simultaneously, design an adaptive process noise covariance matrix. It monitors the variance of IMU acceleration in real time. When an energy peak with a frequency close to the target frequency is detected, it automatically increases the noise covariance of the corresponding state dimension, making the filter more reliant on external observations (visual / UWB) to correct the sway estimation, thereby achieving decoupling compensation for low-frequency large sway and high-frequency structural vibration.

[0081] In step S5, the system switches to dynamic trajectory planning and active compensation control mode to perform fine approximation, specifically including:

[0082] S501. Based on the error vector between the current real-time attitude and the target pose parameters, construct a nonlinear model to predict the control objective function;

[0083] S502. Under the conditions of satisfying the maximum speed, maximum acceleration and anti-collision constraints of the crane, the control command sequence in the future time domain is optimized by rolling.

[0084] S503: The optimized control commands are decomposed into the crane's luffing, slewing, and hoisting actions, as well as the attitude fine-tuning commands of the inclined column itself, to achieve millimeter-level trajectory tracking;

[0085] Among them, fine approximation is 1m from the docking surface.

[0086] In step S6, the specific process of activating the quick-release interface self-locking includes:

[0087] S601. When the vision system detects that the distance between the magnetic guide cone pin at the end of the inclined column segment and the reserved hole in the foundation segment is less than 50mm, the vision servo closed-loop guidance is activated.

[0088] S602. When the contact sensor detects a physical contact signal, it immediately triggers the pneumatic clamping ring to inflate and lock, generating an axial locking force greater than 50kN.

[0089] In step S6, RFID verification specifically includes:

[0090] S603. Read the identification information of the RFID chip embedded at the end of the inclined column segment.

[0091] S604. Compare the installation sequence information with the preset information in the local BIM model.

[0092] S605: If they match, send an unlock command; if they do not match, issue an audible and visual alarm and lock the hook.

[0093] In step 7, the specific implementation steps for total station verification include:

[0094] S701. After the entire inclined column is assembled, select at least three characteristic control points at the top, bottom and middle of the inclined column for total station coordinate measurement.

[0095] S702. Import the measured coordinate data into the data processing software and calculate the cumulative length deviation and overall straightness deviation between each segment.

[0096] S703. If the total length error exceeds ±2mm or the straightness deviation exceeds the design allowable value, the local fine-tuning program is initiated. The hydraulic jacking device is used to release stress and correct the position of specific segments until the accuracy requirements are met.

[0097] By integrating multi-scale sensing fusion of "UWB wide-area positioning + IMU real-time inertial measurement + visual end-point guidance" and combining it with model predictive control, the traditional centimeter-level hoisting error is controlled to a total length deviation of ≤±2mm, meeting the stringent requirements of precision splicing of ultra-large inclined columns, achieving "millimeter-level" extreme precision, and improving the alignment accuracy of inclined columns with large inclination.

[0098] By using RFID identity verification and BIM sequence comparison to form a hard logic lock, the risk of rework caused by incorrect component installation sequence is fundamentally eliminated; and by using a pneumatic clamping ring to provide an axial locking force of more than 50kN, combined with contact sensor triggering, the stability of the connection at the moment of connection is ensured and slippage accidents are prevented.

[0099] In summary, this solution, by deeply integrating multi-source sensing technology, advanced control theory, and BIM digitalization methods, successfully transforms the hoisting of inclined columns with large slopes from an "experience-dependent" to a "data-driven" approach, significantly improving construction safety and efficiency while ensuring extremely high precision.

[0100] Example 2:

[0101] To facilitate the implementation of the aforementioned accurate and rapid alignment method, a system for the precise and rapid alignment of inclined columns with large slopes is proposed. This system includes:

[0102] Global coordinate construction module: used to deploy UWB base station clusters and visual calibration boards and unify spatial reference;

[0103] Digital twin modeling module: used to scan basic segments, reconstruct local BIM models, and solve target pose;

[0104] Intelligent sensing terminal: including reflective markers attached to the inclined column, IMU module and RFID chip;

[0105] Attitude estimation and control module: Built-in attitude estimator algorithm and dynamic trajectory planning and active compensation control mode controller for real-time attitude estimation and dynamic trajectory planning;

[0106] Actuators: including a crane with servo drive function and a quick-release interface with magnetic guide and pneumatic locking function at the end;

[0107] Quality verification module: used for final total station data acquisition and error analysis.

[0108] The quick-release interface also integrates a pressure sensor and a displacement sensor to monitor the contact force and insertion depth during the locking process in real time, and feeds the data back to the attitude calculation and control module to form a closed-loop control.

[0109] The mathematical model of the attitude estimator algorithm is as follows:

[0110] A tightly coupled architecture is adopted, which directly uses the camera's raw pixel observations (feature point coordinates) and the UWB's raw range values ​​(Range) as observation inputs, and performs joint optimization with the IMU's pre-integrated data;

[0111] First, define the system at time... state vector .

[0112] in, : The position of the inclined column in the global coordinate system.

[0113] : The velocity of the inclined column.

[0114] : A unit quaternion representing attitude.

[0115] The zero bias of the accelerometer drifts slowly over time.

[0116] The zero bias of the gyroscope drifts slowly over time.

[0117] in Represents a three-dimensional real vector space. This represents a four-dimensional real vector space.

[0118] Secondly, state extrapolation is performed using high-frequency data from the IMU:

[0119] Input IMU measured specific force and angular velocity :

[0120] De-biasing treatment: ,

[0121] In the formula, This indicates the ratio after removing the bias. This indicates the zero bias of the accelerometer. For accelerometer noise, Indicates the angular velocity after deflection. This indicates the zero bias of the gyroscope. This is gyroscope noise.

[0122] Dynamic equations (continuous time): , ,

[0123] In the formula, For the rate of change of position, For speed, This involves converting a quaternion into a rotation matrix, where g is the acceleration due to gravity. For matrix transpose, It is a differential equation for quaternions.

[0124] Discretization derivation: Using the Runge-Kutta method or the median integral method, based on the time interval... From the current moment Proceeding to the next moment Obtain the nominal state .

[0125] Covariance prediction: Simultaneously extrapolating the covariance matrix of the error state , reflecting the uncertainty of the current estimate, in the formula Indicates the next moment.

[0126] Finally, when new data is received from external sensors (UWB or vision), the following update steps are performed:

[0127] Scenario A: UWB observation update:

[0128] Observations: Labels on the inclined column up to the first Observation distance of each base station ;

[0129] Measured value: Label on the inclined column up to the first Measured distance of each base station ;

[0130] Predicted value: based on the current estimated location Calculate the theoretical distance ;

[0131] For the first The coordinates of each base station;

[0132] Residual: ;

[0133] Kalman gain: ;

[0134] In the formula: This represents the uncertainty of the system's prediction. Represents the observation matrix. This represents the measurement noise covariance. Represents the observation matrix The transpose of .

[0135] Correction: Add the error back to the nominal state and reset the covariance, where Let K represent the state error vector, K represent the Kalman gain, and r represent the Kalman gain.

[0136] Scene B: Visual observation update:

[0137] Observations: Feature points on the image pixel coordinates .

[0138] Projection model: Projecting the world coordinates of known marker points on the inclined cylinder. Based on the current estimated position Transform to the camera coordinate system.

[0139] Then project onto the pixel plane: ,

[0140] In the formula, For projection function, Representing feature points on an image Predicted pixel coordinates on the image, superscript Indicates prediction, This represents the rotation matrix from the world coordinate system to the camera coordinate system. Let R(.) be the quaternion of the current estimated tilted cylinder attitude of the system, and let R(.) denote the function that converts the quaternion into a rotation matrix. Represents matrix transpose. Mark the point Known three-dimensional coordinates in the world coordinate system The system's current estimated camera position coordinates.

[0141] This invention utilizes a UWB base station cluster, a visual calibration board, and a hand-eye calibration algorithm to eliminate coordinate system deviations of multiple sensors in blind spots, establish a globally unified coordinate system centered on the calibration board, and ensure that all measurement data are aligned under the same spatial reference. A local BIM model is generated by scanning the completed foundation segments and then ICP-registered with the theoretical BIM model.

[0142] By calculating the actual deviation to correct the theoretical docking coordinates, a precise target pose containing position and attitude is generated, solving the problem of "construction according to the drawings is incorrect" caused by foundation settlement or construction errors.

[0143] During the hoisting process, an augmented extended Kalman filter is constructed. Instead of relying on traditional position and velocity estimation, the sling swing mode is incorporated as a state variable into the calculation. A damped simple harmonic oscillation model is introduced to describe the swaying of the inclined column under the long sling. At the same time, through adaptive noise covariance technology, IMU, UWB and visual data are intelligently fused to achieve accurate decoupling and real-time compensation of low-frequency oscillation and high-frequency vibration, and output a realistic six-degree-of-freedom attitude.

[0144] Within a precise approximation range of 1 meter from the docking surface, the system switches to nonlinear model predictive control mode. Based on the error between the current attitude and the target pose, combined with the crane's physical limits (speed, acceleration) and anti-collision constraints, the optimal control command is optimized by rolling and decomposed into the crane's large movements (amplitude change, slewing) and the column's fine-tuning movements, achieving millimeter-level trajectory tracking.

[0145] Once the vision-guided magnetic cone pin enters a 50mm range, the vision servo closed loop is activated until physical contact is made, triggering the pneumatic clamping ring to generate a 50kN locking force for mechanical locking. Subsequently, the identification information is read via an RFID chip and compared with the installation sequence in the BIM model. Only if the information matches is the hook allowed to be released, preventing the wrong component from being lifted, thus achieving a dual closed loop of physical connection and digital information.

[0146] After the entire inclined column is assembled, the characteristic points are measured and verified using a total station. If the total length error or straightness exceeds the allowable range (≤±2mm), the hydraulic jacking device is activated to release local stress and make corrections, ensuring the millimeter-level accuracy of the final product.

[0147] Components not described in detail in this article are existing technologies.

[0148] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for splitting and accurately and quickly aligning a large slope inclined column, characterized in that, Comprising the following steps: Step S1: On-site deployment of UWB base station group and visual calibration board, complete global coordinate system one; Step S2: Scan the completed basic section, reconstruct the local BIM model and export the target pose parameters; Step S3: Transport the prefabricated inclined column section to the site, paste the reflective markers, and load the IMU module; Step S4: Slowly lift after the crane hook is hooked, and simultaneously start the pose estimator; Step S5: Switch to dynamic trajectory planning and active compensation control mode, and perform fine approximation; Step S6: Activate the quick disassembly interface self-locking after reaching the preset threshold, and release the hook after confirming that the RFID verification is passed; Step S7: Repeat the above process until the entire inclined column is assembled, and finally use the total station instrument to check the total length error ≤±2mm.

2. The method of claim 1, wherein, In step S1, the specific process of completing the global coordinate system one includes: S101, Use the visual calibration board as a common reference to collect the absolute coordinates of the UWB base station group in the global coordinate system and the relative coordinates in the visual camera coordinate system; S102, Calculate the rotation and translation transformation matrix between the UWB coordinate system and the visual coordinate system using the hand-eye calibration algorithm; S103, Define the center point of the visual calibration board as the origin of the global coordinate system, and map all sensor data to the global coordinate system to eliminate the spatial reference deviation of multi-source heterogeneous data.

3. The method of claim 1, wherein the method is characterized by, In step S2, the specific process of reconstructing the local BIM model and exporting the target pose parameters includes: S201, Denoise, register and reconstruct the surface of the collected point cloud data to generate a high-precision three-dimensional mesh model of the basic section; S202, Perform iterative nearest point registration between the generated three-dimensional mesh model and the theoretical BIM model in the design stage, and calculate the deviation matrix between the actual installation state and the design state; S203, correcting the coordinate values of the interface based on the bias matrix correction theory to generate a target pose parameter set containing a position vector and a pose quaternion and send it to the control terminal.

4. The method of claim 1, wherein, In step S4, the running logic of the pose estimator specifically includes: S401, Use the angular velocity and acceleration output by the IMU module as the prediction input, and use the absolute position output by the UWB base station group and the relative pose calculated by the visual system capturing the reflective markers as the observation input; S402, Construct an extended Kalman filter state equation to dynamically compensate for high-frequency vibration, wind disturbance and sling elastic deformation of the inclined column during hoisting; S403, Real-time output of six-degree-of-freedom attitude data as feedback signal for crane motion control.

5. The method of claim 1, wherein, In step S5, switch to dynamic trajectory planning and active compensation control mode, and perform fine approximation, which specifically includes: S501, Based on the error vector of the current real-time pose and the target pose parameters, construct a nonlinear model predictive control objective function; S502, Under the conditions of satisfying the crane speed, acceleration and anti-collision constraints, roll the control instruction sequence in the future time domain; S503, Decompose the optimized control instruction into amplitude, rotation, lifting actions of the crane and attitude fine-tuning instructions of the inclined column itself to realize millimeter-level trajectory tracking; Wherein, the fine approximation is 1m away from the docking surface.

6. The method of claim 1, wherein, In step S6, the specific process of activating the quick disassembly interface self-locking includes: S601, When the visual system detects that the distance between the magnetic attraction guide cone pin at the end of the inclined column section and the reserved hole position of the basic section is less than 50mm, start the visual servo closed-loop guidance; S602, when the contact sensor detects a physical contact signal, trigger the pneumatic clamping ring inflation locking immediately, generate an axial locking force greater than 50kN.

7. The method of claim 1, wherein the method is a method of splitting and precisely and rapidly aligning a large slope bevel, characterized by, In step S6, the RFID verification specifically includes: S603, read the RFID chip identity information embedded in the end of the inclined column section, S604, compare with the preset installation sequence information in the local BIM model, S605, if consistent, send the unlocking instruction, if not consistent, issue an audible and visual alarm and lock the hook.

8. The method of claim 1, wherein, In step 7, the specific implementation steps of total station review include: S701, after the assembly of the whole inclined column is completed, select at least three feature control points at the top, bottom and middle of the inclined column for total station coordinate measurement; S702, import the measured coordinate data into the data processing software, calculate the cumulative length deviation and overall straightness deviation between each section; S703, if the total length error exceeds ±2mm or the straightness deviation exceeds the design allowable value, start the local fine tuning program, release the stress and correct the position of the specific section through the hydraulic jacking device until the accuracy requirement is met.

Images